COMPUTATION OF TOTAL TRU ACTIVITY EXCISED IN THE
KICKAPOO AREA OF SALLY
DOE/ERSP TECH NOTE NO. 10.0

DATED: 28 July 1978

AUTHOR: M. Barnes, DRI
Two different computations of total TRU activity (in curies) removed from Kickapoo were made.

One was based on soil samples taken from each individual truckload of soil which were
gamma-scanned for 241 am activity. The other was based on IMP readings on the surface, taken
before, during, and after the excision process. Both estimates required knowledge of certain
information from outside sources; unfortunately, the information was not always consistent or
accurate. Therefore, this note will explain in detail only the methods and mathematics used in
deriving the estimates.
The actual estimates will be shown for each of the various sets of information from outside sources,

since ERSP is not in a position to judge the validity of such information.
Estimates Using Truck Soil Samples

Each truck was soil sampled using one or both of two different methods. Originally, three samples
were taken from the top of each truck after loading and composited to form "top" samples. This
method has obvious statistical drawbacks, including being biased high as an estimate of the truck
average. Later, a sample was taken from each scoop going into the trucks, and the samples from all
scoops for each truckload were composited to form "mixed" samples. This method, while not as
biased as the original one, still is biased high. Bias is present in both methods due to the fact that
the dispersion variance* of soil samples within a truck increases with average concentration.

Thus

high values should be, but are not, given less weight in estimating the average concentration in a

truck. (No data are available to compute the proper weights.)

The two methods were compared for the thirty truckloads for which both types of samples were
taken. The mean of the top samples was 31.7 pCi/g TRU, with a sample standard deviation of 29.8.
value.

However, 20 of the 30 pairs had a higher top sample value than mixed sample

A sign test was performed to test the hypothesis that the two types of samples came from

distributions having the same median. This hypothesis can be rejected at the 95% confidence level,
i.e., the median of the top sample distribution is significantly higher than the mixed sample median.

ry

The mixed samples had similar results, with a mean of 25.8 pCi/g TRU and sample standard

deviation of 32.3.

Estimates of total activity were made by multiplying the cubic yards held by a truck by the
concentration in each sample from that truck and summing the cubie yards for total volume and the
products for total activity. Mixed sample results were used whenever available. Truck sizes (by
truck number, which was the soil sample identifier) were obtained from S-3, 84th Engineer
Battalion. The nominal cubie yardages for each truck size were also provided by S-3, 84th

Engineers, but two different values were given at different times, as follows:
Date of Yardage
Information

17 July

Nominal Cubie Yards Per Truck

o Ton

10 Ton

20 Ton

Total

Volume

Total

Curies

1978

3

3

12

5500 cu. yds.

0.95

22 July 1978

3

3

10

4500 cu. yds.

0.77

The truck sample data were 241 am by gamma sean, and a fixed ratio of 6.16 was used to convert to

TRU concentrations.

*Dispersion variance of soil samples within a truck defined as the variance of the distribution of

concentration values from every possible soil sample within each truck.

B-10-1

wa

Therefore, following this comparison experiment, all samples taken were of the mixed type.

Select target paragraph3